2.1 Electric field method

The scheme of the designed electric field method for suppressing the secondary electrons is shown in Fig. 2. A ceramic ring is used to isolate the target from the extraction-accelerating electrode. A resistance (R) is used to connect the extraction-accelerating electrode to the target. When the deuterium current bombards the target and flows through the resistance to the extraction-accelerating electrode, a bias voltage is generated between the target and the extraction-accelerating electrode. The bias voltage (V_b) of the target relative to the extraction-accelerating electrode can be estimated with the following equation,

Fig. 2Full-screen View

Scheme of the electrostatic suppression.

where, I_d is the deuterium ion current on the target. A current of secondary electrons, which is formed by the secondary electrons hitting the extraction-accelerating electrode, also affects the calculation of V_b through the resistance R to the target. The aim of this experiment is to obtain the threshold suppression voltage at which the secondary electrons are effectively suppressed. In this case, the secondary electron current can be negligible.

According to Eq.(1), when I_d is constant, V_b increases linearly with increasing R. The secondary electrons should be completely suppressed when V_b is large enough. To study the performance of the secondary electron suppression of the electric field method shown in Fig. 2, experimental measurements of the high voltage load current (I_load) with the accelerating voltage (V_ex) are taken under the same arc current (I_arc) of the ion source and various resistors (R = 0 kΩ, R = 10.2 kΩ, R = 51.1 kΩ, R = 102 kΩ, R = 133.2 kΩ, R = 200 kΩ, R = 400 kΩ, and R = 500 kΩ). The high voltage load current I_load is the read back of the high voltage power supply to the given high voltage. These resistance values are obtained by a combination of different metal film resistors. The rated power of a single metal film resistor is 2 W. The current I_load should be the sum of the extraction deuterium current (I_d), the secondary electron current (I_e), and the leakage current of high voltage (I_leak):

The experimental results of I_leak, with I_d = 0 mA, show that I_leak will not be greater than 0.1 mA. even at accelerating voltage of 150 kV, thus, I_leak can be ignored. Therefore, I_load can be approximately calculated as the sum of I_d and I_e:

Fig. 3 shows the experimental results of I_load as a function of the accelerating voltage (V_ex) for various resistances (R) under the same arc current (I_arc) of the ion source. As shown in Fig. 3, for R = 0 kΩ (V_b = 0 V), I_load increases rapidly with increasing V_ex and, when V_ex = 120 kV, I_load is up to 9.2 mA. As R increases, the increase in I_load becomes slower with increasing V_ex and, at the same V_ex, I_load decreases rapidly. This means that more and more secondary electrons are suppressed and, when R = 0 kΩ (V_b = 0 V), a large part of I_e is in part of the I_load. When R ≥102 kΩ and V_ex ≥ 80 kV, I_load becomes a constant of 2 mA. This suggest that the secondary electrons have been completely suppressed, and I_load is equal to I_d. In other words, the effective secondary electrons suppression can be achieved when the bias voltage between the target and the extraction-accelerating electrode is greater than or equal to 204 V (the product of R = 102 kΩ and I_d = 2 mA from Eq. (1)).

Fig. 3Full-screen View

The load current (I_load) as a function of the accelerating voltage (V_ex) for different resistances (R), with I_arc = 0.5 A.

2.2 Magnetic field method

The designed magnetic field method for suppressing secondary electrons is shown in Fig. 4a. A pair of magnets are aligned with opposite poles facing each other to produce a magnetic field parallel to the target surface. The target is connected to the extraction-accelerating electrode through a wire. The Sm₂Co₁₇ permanent magnet is selected because of its high remanent flux density and thermostability. Under three different permanent magnet sizes (Φ 20 × 5 mm, Φ 20 × 10 mm, and Φ 20 × 12 mm), the magnetic field distributions along the center axis of the magnets (the red line in Fig. 4a) are measured with a Hall magnetometer. The results are shown in Fig. 4b. The magnetic field strength at the center of the target surface (X = 0 mm) is approximately 76 Gauss for Φ 20 × 5 mm, 131 Gauss for Φ 20 × 10 mm, and 150 Gauss for Φ 20 × 12 mm.

Fig. 4Full-screen View

(Color online) Placement of the permanent magnets (a); magnetic field distributions along the center axis of the magnets for different magnet sizes (b).

To study the performance of secondary electron suppression of the magnetic field method shown in Fig. 4a, the load current (I_load) as a function of the accelerating voltage (V_ex) were measured with and without a magnet (three different sizes) under the same arc current (I_arc) of the ion source. The experimental results are shown in Fig. 5. It can be seen that I_load is significantly lower with the Sm₂Co₁₇ magnets. The secondary electrons can be suppressed by the designed magnetic field method. When the magnet size is Φ 20 × 5 mm, the secondary electrons are not completely suppressed. The effects of secondary electron suppression with the magnets of size Φ 20 × 10 mm and Φ 20 × 12 mm are superior than those with the Φ 20 × 5 mm magnets.

Fig. 5Full-screen View

Load current (I_load) as a function of accelerating voltage (V_ex) with and without magnets, with I_arc = 0.5 A.

2.3 Comparison of the electrostatic and magnetic suppressions

Both the electric field and magnetic field methods have affected the suppression of secondary electrons. To compare their performance in secondary electrons suppression, the dependence of I_load on V_ex for different arc currents (I_arc) of the ion source were measured with the electric field method with R = 500 kΩ and the magnetic field method with the magnet of size Φ 20 × 12 mm. During the experiment, the accelerating voltage (V_ex) was increased over time from 15 kV to 120 kV, in steps of 5kV, and the arc current (I_arc) was increased over time from 0.4 A to 0.8 A, in steps of 0.2 A. The experimental results are shown in Fig. 6. As shown in the figure, for different I_arc, the I_load gradually reaches a stable value with increasing V_ex in both methods. This confirms that both methods can effectively suppress secondary electrons. However, the I_load with the magnetic field method is significantly higher than that with the electric field method, while the I_arc of the ion source is higher in the former. A higher arc current (I_arc) means a bigger deuterium ion beam, that is, the suppression effect of the magnetic field will become worse as the deuterium ion beam increases. There are two possible reasons for this: one is the space charge effect, which can occur in the vicinity of the target. and the other is the demagnetization of the permanent magnet due to ion bombardment and heat conduction from the target.

Fig. 6Full-screen View

Load current (I_load) as a function of the accelerating voltage (V_ex) with different I_arc for electrostatic suppression with R = 500 kΩ and magnetic suppression with the magnet of size Φ 20 mm (diameter) × 12 mm (thickness).

In summary, the designed electric field method is superior to the magnetic field method for secondary electrons suppression. Therefore, the electric field method will be employed in our CDDNG. For a better suppression effect in a lower deuterium ion current, the resistance should be 500 kΩ. The resistance is located in the transformer oil of the high voltage insulator assembly to achieve good cooling. After the implementation of the designed electric field method, the CDDNG was continuously operated for more than 3 hours without high voltage breakdown. When measures to suppress secondary electrons are not taken, high voltage breakdown occurs frequently, especially when the load current is close to 10 mA. Furthermore, the dose of X rays is reduced by about 10 times.

2.4 Estimation of the secondary electron emission coefficient

In this section, the secondary electron emission coefficient for deuterium ions incident on molybdenum is estimated based on the experimental data of the electrostatic suppression shown in Fig. 3. The secondary electron emission coefficient should be the ratio of the number of secondary electrons ejected from the target to the number of incident deuterium ions. It is assumed that the secondary electrons are not suppressed at all with R = 0 kΩ, and are completely suppressed with R = 500 kΩ and V_ex> 80 kV. According to the experimental data in Fig. 3, the secondary electron emission coefficient λ of deuterium beam incident on molybdenum can be estimated by the following equation:

where, E_d, which is approximately equal to the product of V_ex and the charge of deuterium ion, is the energy of the deuterium ion, I_load(0, $E_{\text{d}}$ ₎ is the high voltage load current for R= 0 kΩ, and I_load(500, $E_{\text{d}}$ ₎ is the high voltage load current for R= 500 kΩ.

However, some secondary electrons can be suppressed even if V_b =0 V (i.e. R= 0 kΩ), owing to the target being installed inside the cavity of the cylindrical extraction-accelerating electrode (see Fig. 2). In other words, I_load(0, $E_{\text{d}}$ ₎ - I_load(500, $E_{\text{d}}$ ₎, which does not include the suppressed secondary electron current, is only the back-accelerated secondary electron current for R = 0 kΩ. If the coefficient λ is directly estimated according to the experimental data in Fig. 3 and equation (4), it will be underestimated. If the ratio(η) of the back-accelerated secondary electrons to the total secondary electrons ejected from the target can be found, equation (4) can be modified to:

It is quite difficult to measure the value of η. Therefore, a simulation investigation on the transport of the secondary electrons was carried out by a PIC code ^[16]. In the simulation, we skip the complex physical process of the secondary electrons produced by the deuterium ion interaction with the target ^[17], because electrons are produced as long as the deuterium ions are incident on the target surface. It is assumed that a deuterium ion produces three secondary electrons. The angular distribution of electrons ejected is set to be a cosine distribution. After contact with the target, the deuterium ions are 'killed', which means they disappear in the subsequent simulation, irrespective of the cascade effect of the secondary electrons, which means that the secondary electrons disappear directly after hitting any electrode and there is no secondary particle production. The current of deuterium ions is 2 mA.

A typical simulation image of the space distribution of the secondary electrons is shown in Fig. 7a, under V_ex = 120 kV and R = 0 kΩ. Fig. 7b shows the potential distribution between the extraction-accelerating electrode and the target. The value of η can be calculated by counting the number of secondary electrons on each electrode. The calculated results of η are shown in Fig. 7c, with V_ex increasing from 80 kV to 120 kV in steps of 10 kV. The value of η as a function of V_ex shows a good linear relationship. The linear fitting values of η are used to calculate the coefficient λ. Fig. 8 shows the calculated results of the coefficient λ according to equation (5) with the correction of the ratio η. It can be seen that λ increases with increasing deuterium beam energy. In the energy range of 80 to 120 keV, the value of λ is between 5–5.5 with a deuterium ion glancing incidence of 45°.

Fig. 7Full-screen View

(Color online) (a) Space distribution of the secondary electrons for V_ex = 120 kV and R = 0 kΩ; (b) potential distribution between the extraction-accelerating electrode and the target, where T and E denote the target and the extraction-accelerating electrode, respectively; (c) ratio of the back-accelerated secondary electrons to the total secondary electrons (η) ejected from the target as a function of V_ex, for R = 0 kΩ.

Fig. 8Full-screen View

Secondary electron emission coefficient (λ) for deuterium beam incident on molybdenum as a function of the deuterium energy (E_d).

The secondary electron emission coefficient on the metal target bombarded by deuterium ion has been widely investigated in the past decades ^[18-20]. However, most of these studies were carried out under the condition of normal incidence. According to references ^[21,22], the secondary electron emission coefficient with glancing incidence of 45° (λ_45°) can be calculated by the following equation,

where λ_⊥ is the secondary electron emission coefficient under normal incidence. For comparison with the previous data, we convert λ_45° into λ_⊥ according to equation (6). Fig. 8 shows the calculated results. The value of coefficient λ_⊥ is between 3.5–3.9 at energies of 80–120 keV. The experimental data from reference ^[20] for the deuterium ions (D₁⁺) normally incident on molybdenum are also shown in Fig. 8. It can be seen that λ_45° is bigger than λ_⊥. This is because the glancing incidence leads to a shallower penetration and it makes it easier for the generated carriers to exit the target surface. Furthermore, our estimated results (λ_⊥) are significantly greater than the experimental data from reference ^[17]. The reason is that the mixed deuterium ion beam (D₁⁺, D₂⁺, and D₃⁺) is used in our experiment. The previous experimental results show that the coefficient (λ_⊥) for D₂⁺ ions is about two times that of equal-energy D₁⁺ ions ^[23]. According to our previous experimental results ^[24], the proportion of D₁⁺ ions in the mixed deuterium ion beam (D₁⁺, D₂⁺, and D₃⁺) is less than 10% under the condition of low arc current (such as I_arc< 1 A) for the duoplasmatron ion source.